An Early Appointment to Outpatient Cardiac Rehabilitation at Hospital Discharge Improves Attendance at OrientationClinical Perspective
A Randomized, Single-Blind, Controlled Trial
Background—Outpatient cardiac rehabilitation (CR) decreases mortality rates but is underutilized. Current median time from hospital discharge to enrollment is 35 days. We hypothesized that an appointment within 10 days would improve attendance at CR orientation.
Methods and Results—At hospital discharge, 148 patients with a nonsurgical qualifying diagnosis for CR were randomized to receive a CR orientation appointment either within 10 days (early) or at 35 days (standard). The primary end point was attendance at CR orientation. Secondary outcome measures were attendance at ≥1 exercise session, the total number of exercise sessions attended, completion of CR, and change in exercise training workload while in CR. Average age was 60±12 years; 56% of participants were male and 49% were black, with balanced baseline characteristics between groups. Median time (95% confidence interval) to orientation was 8.5 (7–13) versus 42 (35 to NA [not applicable]) days for the early and standard appointment groups, respectively (P<0.001). Attendance rates at the orientation session were 77% (57/74) versus 59% (44/74) in the early and standard appointment groups, respectively, which demonstrates a significant 18% absolute and 56% relative improvement (relative risk, 1.56; 95% confidence interval, 1.03–2.37; P=0.022). The number needed to treat was 5.7. There was no difference (P>0.05) in any of the secondary outcome measures, but statistical power for these end points was low. Safety analysis demonstrated no difference between groups in CR-related adverse events.
Conclusions—Early appointments for CR significantly improve attendance at orientation. This simple technique could potentially increase initial CR participation nationwide.
Cardiac rehabilitation (CR) is known to decrease the subsequent risk for total mortality, cardiovascular mortality, and cardiovascular morbidity after myocardial infarction (MI),1,2 percutaneous coronary intervention (PCI),3 and coronary artery bypass graft surgery4 but is highly underutilized. Nationally, <30% of patients participate in CR after an MI,5,6 and there has been only minimal improvement in the past 20 years.7 Improving CR participation rates has recently been the subject of both an American Heart Association Presidential Advisory8 and a Scientific Advisory9 and is an ongoing national priority for improvement.10
Clinical Perspective on p 355
National guidelines suggest the ideal time interval between hospital discharge to CR enrollment should be between 1 and 4 weeks, depending on type of index event.11,12 However, the time from hospital discharge to CR enrollment currently averages ≈35 days13,14 and represents a modifiable factor that could potentially affect participation. Preliminary retrospective studies suggest that for every day of delay between hospital discharge and CR attendance, there is an associated 1% decrease in participation.14 The reasons for this delay are unclear but may relate to the need to obtain insurance preauthorization, difficulty in arranging patient appointments after hospital discharge, program capacity restraints, or residual concerns about the safety of early exercise after MI. Recently, however, the safety of early exercise after MI or PCI has been demonstrated15–17 and should no longer be a significant consideration for patients who are stable.
We therefore undertook a randomized, single-blind, controlled trial to test the hypothesis that an early (within 10 days) versus standard (at 35 days) scheduled appointment made at hospital discharge for a free-of-charge orientation class for CR would result in significantly greater attendance in the class. Secondary outcomes included participation rates for attending ≥1 CR exercise session, total number of exercise sessions attended, changes in exercise training workload over time while in CR, CR completion rates, and the exercise-related and overall clinical safety of such a referral strategy.
Inpatients at Henry Ford Hospital in Detroit, MI, were recruited to participate between February and November 2011, with follow-up occurring through May 2012. Patients were eligible to participate if they had a qualifying diagnosis for referral to CR, were >18 years of age, granted access to their medical record for research, and gave written informed consent. Qualifying diagnoses included having an MI, PCI, or angina with an ischemic stress ECG, stress echocardiogram, or stress myocardial perfusion imaging study. For this trial, although eligible for CR, patients who had undergone recent coronary artery bypass grafting, valve surgery, or cardiac transplantation were excluded.
Additional exclusion criteria included patient refusal to attend CR, plans to attend CR outside of the Henry Ford Health System, current or previous enrollment into CR within the prior 6 months, the presence of moderate or severe dementia, unstable psychiatric condition, severe peripheral vascular disease that precluded exercise, uncorrected severe aortic stenosis, uncorrected severe mitral stenosis, presence of a left ventricular assist device, discharge to a nursing home or rehabilitation center, planned future medical care outside of the Henry Ford Health System, or safety concerns that precluded exercise per the discretion of the treating/referring physician. We did not exclude underinsured or uninsured patients because these patients could attend CR orientation without cost, and a self-pay option was available for those who might choose to participate in CR exercise sessions.
Randomization and Blinding
Eligible patients were identified by their treating resident physician. Patients were then approached by the study personnel, who consisted of 8 residents in internal medicine and 2 cardiology fellows. Patients were informed of the referral to CR, strongly encouraged to attend CR, provided a choice of rehabilitation site (among 5 Henry Ford Health System CR centers in the greater Detroit area), given literature about CR, and given a map with contact information for their chosen site. After signing informed consent, patients completed a 10-question survey based on the transtheoretical model (stages of change) of behavior change. Patients were told this survey was the purpose of the study and were thus blinded to the main study hypothesis to avoid the Hawthorne effect. At study end, a full disclosure letter revealing this blinding technique was sent to all patients. This study design was reviewed and approved by the Henry Ford Institutional Review Board.
Patients were then randomized to receive either an early (within 10 days) or a standard-of-care (35 days) appointment for the CR orientation class.13,14 Each of the 5 CR sites offered a specified weekly orientation session, and information about a patient’s specific appointment day and time was delivered to them by a study representative. Residents and fellows were instructed to minimize patient contact after randomization: They briefly noted the date, time, and location for the first appointment in CR; collected completed surveys; provided a program brochure; and answered simple residual questions about CR in general. These tasks typically required less than 2 to 3 minutes of contact time after randomization, but residents and fellows were not blinded by treatment group.
Patients randomized to the standard-of-care group were not forbidden from attending an earlier orientation class. Although they were given an official 5-week appointment, a routine referral was also generated through normal hospital referral patterns and forms. This option allowed the treating cardiologist or other provider to expedite enrollment into CR if desired or clinically necessary, thus matching the current standard of care.
Randomization was performed in a 1:1 pattern with 2 blocks. The first block was used exclusively during daytime hours and was kept by a receptionist in CR who was available via central telephone. The second block was kept by the principal investigator (Q.R.P.), who was available by pager for weekend and evening randomization. Sequence generation was created by the senior author (S.J.K.) using a computerized random number generator. All allocation cards were kept in opaque sequential sealed envelopes until the time of patient randomization. Neither the senior author nor the secretary recruited patients, but the principal investigator did recruit patients.
After hospital discharge, to facilitate the collection of pertinent social, demographic and health insurance information, all patients received a single phone call before their scheduled appointment for orientation into CR. Because CR staff needed open access to the medical record to care for patients, it was not possible to blind them to treatment group. If a patient failed to attend the specified orientation appointment (regardless of group assignment), CR staff contacted the patient per standard practice procedures until either the patient refused CR, the patient attended orientation, or further efforts to enroll the patient were deemed futile by CR staff. No study investigators were involved in these interactions. On enrollment, CR staff treated both groups of patients equally in terms of evaluation, exercise testing and prescription, follow-up, and clinical care.
The primary outcome was attendance at the orientation class for CR, which all patients must attend before returning for their first exercise session. Any patient lost to follow-up was considered a nonattendee. In addition, any patient who attended an orientation session >5 weeks from their originally scheduled appointment date was considered a nonattendee. Beyond 5 weeks, we believe that factors other than the specific randomized appointment time (eg, follow-up with a cardiologist or primary doctor) may affect a patient’s decision to attend CR rather than an already missed specific appointment time that was more than a month in the past.
Secondary outcomes were as follows: (1) Time from hospital discharge to attendance at CR orientation; (2) time difference between the scheduled and actual date of orientation; (3) attendance at ≥1 exercise and education session of CR; (4) total number of exercise sessions attended; (5) completion of CR, defined as either meeting the prespecified number of sessions at initiation of CR or attending ≥12 sessions; and (6) for patients who attended ≥2 exercise sessions, the change in average exercise work load18 (expressed as estimated metabolic equivalents [METs]) while training on the treadmill. CR staff monitored and recorded each of these outcomes per routine clinical practice.
Exercise-related safety events (eg, falls, syncope, angina, and emergency department visits) were determined by review of the medical record by 2 medical residents (M.M. and J.S.B.). An event was considered to be related to CR if it occurred either during or within 3 hours after a CR exercise session. In addition, non–exercise-related clinical events that occurred within the first 6 months after randomization were identified from the medical record. Events considered were recurrent MI, coronary artery bypass graft surgery, unplanned PCI, stroke, or hospitalization for unstable angina.
On the basis of an estimated absolute effect size of 20%,15 a control group attendance of 50% (baseline enrollment rates at our institution were 44% [unpublished data]), an intervention attendance of 70%,19 α=0.05, and β=0.80, we estimated that 206 patients (103 per group) would be needed. No interim analysis was originally planned, and no prespecified stopping boundaries were determined. However, because of the relocation of the trial’s principal investigator (Q.R.P.), a single interim analysis was performed (by J.K.E.). For this analysis, a patient was considered to have attended CR if they attended orientation within 3 weeks of their scheduled appointment instead of the standard 5-week period. The stopping boundary was P<0.05. With 122 patients in the interim analysis, the CR attendance rate was 69% (43/62) for the early appointment group and 45% (27/60) for the standard-of-care group (P=0.01 by Fisher exact test). On November 6, 2011, the cessation of patient recruitment was recommended based on the results from this analysis.
All patients were analyzed in the group to which they were randomized according to intention-to-treat principles. Baseline characteristics for both groups were provided, with descriptive statistics. Proportions of patients attending CR orientation were calculated by group, and baseline differences were compared with 2-group t tests and 2-group Wilcoxon tests as appropriate for continuous variables. For categorical variables, Fisher exact and χ2 testing were used as appropriate. For time-related outcomes, a time-to-event analysis was performed to account for censored data, and groups were compared with log-rank tests. For the safety analysis, we considered only events that occurred within the first 6 months and applied the log-rank test. Any patient lost to follow-up was censored at the time of last contact. P<0.05 was considered statistically significant. All statistical analyses were performed with SAS 9.2 (SAS Institute, Cary, NC).
We recruited and randomized 150 patients. Two patients in the early appointment group withdrew consent, and 148 were included in the final analysis. The enrollment flow diagram is shown in the Figure. Baseline demographic and clinical characteristics of the early appointment and standard-of-care groups are given in Table 1. As shown, at baseline, the 2 groups were generally similar in age, sex, racial diversity, qualifying diagnosis, past medical history, risk factors for coronary artery disease, distance to the rehabilitation center, and insurance status.
All primary and secondary outcomes are shown in Table 2. The early appointment group’s time between discharge and orientation attendance (95% confidence interval) was 8.5 (7–13) days compared with 42 (confidence interval 35 to NA [not applicable]) days for the standard-of-care group (log-rank P<0.001). Ten patients (14%) in the standard-of-care group participated in an orientation session that was earlier than their randomized 5-week appointment time. For the primary outcome, there was an absolute 18% improvement in attendance at CR orientation between the early appointment and standard-of-care group. This corresponds to a 56% relative improvement (relative risk, 1.56; 95% confidence interval, 1.03–2.37; P=0.022). The number needed to treat was 5.7 to have 1 additional patient attend CR orientation.
There were no significant differences in any of the secondary outcomes (Table 2); however, observed power was only 18% for patients attending ≥1 exercise session, with similarly low observed power for other secondary outcomes. Reasons for not attending the orientation class or not exercising are shown in Table 3. The most common reason for not attending orientation was loss of contact with the CR program. The most common reason for not exercising for ≥1 session in CR was financial.
Post hoc analyses revealed that elderly patients (age ≥65 years) were more likely to attend orientation with this strategy than younger patients, with 43 (80%) of 54 elderly patients versus 58 (62%) of 94 younger patients attending orientation (P=0.024). However, there was no apparent difference by race (P=0.81), sex (P=0.29), or MI (with or without ST-segment elevation) as part of the index event (P=0.67).
There were 2 hospitalizations during or within 3 hours after completion of a CR session. One patient was hospitalized for atypical chest pain in the early appointment group, and another was hospitalized for syncopal symptoms in the standard-of-care group. Both were discharged from the hospital uneventfully. There were 2 deaths of previously undiagnosed cancers, 1 in each group, and neither patient attended CR. Safety outcomes in the 6 months after the index event demonstrated 5 non–ST-segment elevation MIs, 1 PCI, and 1 hospitalization for unstable angina in the early appointment group and 1 coronary artery bypass surgery, 1 ST-segment elevation MI, 4 non–ST-segment elevation MIs, 3 PCIs, and 3 hospitalizations for unstable angina in the standard-of-care group (7 versus 12 events for early appointment versus standard of care, log-rank P=0.32).
Consistent with our hypothesis, this randomized, controlled, single-blind trial demonstrated a significant improvement in attendance at the CR orientation class when an early scheduled appointment time (within 10 days) was used versus standard of care (5 weeks.). Importantly, the magnitude of this effect was sizable, with only 5.7 patients needing an early appointment to have 1 additional patient attend orientation. We were unable to demonstrate a difference in participation in CR exercise sessions, program completion rate, number of rehabilitation sessions attended, or change in work load during exercise training. However, we did not design our trial for these secondary end points, and it was underpowered to detect such a difference. Safety analysis demonstrated no difference between groups. Consequently, the only firm conclusion that can be drawn from this trial is that a strategy of early appointments to CR increases attendance at orientation.
The main implication of these findings is that early appointments to CR may represent a new strategy for improving cardiac rehabilitation participation. National enrollment rates are reported to be in the 20% to 35% range.5,6 Given the known substantial mortality benefits of CR after an MI or PCI,1,3 there is a clear imperative to improve participation rates.8 One appealing aspect of an early appointment system, once implemented, is that it should not require significant resources beyond the discharge planning already in place in hospitals. When coupled with other evidence-based techniques such as systematic referral20 and liaison-facilitated referral,21 early appointments to CR could significantly improve CR participation nationwide.
Previously, Russell et al14 estimated that for every 1 day of delay to the initial CR appointment, there is an approximate 1% loss in enrollment. We found a similar effect size (17% observed versus 23% expected, well within the range expected). Similarly, Soga et al15 noted an absolute 21% increase in the rate of exercise continuation among patients who underwent a submaximal treadmill exercise test the day after PCI as part of CR versus those who did not undergo early exercise testing. Parker et al,22 using an early clinic appointment for ST-segment elevation MI patients, also demonstrated significantly improved enrollment in CR. Thus, in several different settings, early contact has been shown to improve CR participation. However, all 3 of the above studies were retrospective and are subject to selection bias, a major limitation overcome by our study design.
However, we were unable to demonstrate a difference in our secondary outcomes. This may be for several reasons. First, our observed power for the secondary end point of attending ≥1 exercise session was only 18%. Still, we observed a nonsignificant 10% favorable difference in the rate of patients beginning to exercise, and other secondary outcomes also moved in the direction we hypothesized. This suggests that at the very least, early appointments did not worsen these secondary outcomes. Second, we did not design the present study for secondary end points, instead choosing to focus on attendance at orientation as the essential first step toward participation in CR. This was done because it allowed us to skip insurance preauthorization as a prerequisite to attendance at orientation, which often causes a 2- to 3-week delay of entry into CR at our institution. Similarly, this design allowed underinsured/uninsured patients to participate, because orientation was without cost, and these patients could then consider self-pay options. Theoretically, if the uninsured patients in the present trial were all insured, 6 versus 2 patients (Table 3) would likely have joined the early appointment versus standard-of-care group, respectively. This would have changed the ≥1 exercise session secondary end point to 49 versus 38 patients, producing an estimated relative risk of 1.30 (95% confidence interval, 0.96–1.72; P=0.066, a statistical trend). Consequently, improved secondary outcomes are still a possible effect of an early appointment and are neither proved nor disproved by this trial. Future trials testing the effects of early appointments should focus on recruiting fully insured patients who do not require preauthorization and should use innovative methods to increase patient retention during orientation.
The reasons this strategy improves CR participation are likely 3-fold. First, a hospitalization for a major coronary event likely represents a time when readiness to make lifestyle changes is at its highest, as has been demonstrated previously among patients who smoke.23 Consequently, capturing this “teachable moment” early in the recovery period likely enhances the chance that patients will begin a regular exercise program, attend CR, or make other lifestyle changes. Second, sending a strong message about the importance of CR is known to affect CR participation.24 However, when patients are given an appointment for “next week” versus “next month,” there is a clear differential message sent about the importance of CR, such that an appointment “next week” implies that CR is integral to full recovery, whereas “next month” implies that CR may be less important. Third, the simple time proximity from hospital discharge makes an early appointment easier to remember and plan for before competing demands distract and dissuade a patient from attending CR.
There are several other findings worthy of comment. First, a strategy of early enrollment is feasible to implement and successfully increases the rapidity of attendance at orientation. Furthermore, patients recently discharged from the hospital are willing and able to keep their outpatient appointment for CR. Second, although we did not demonstrate a significant increase in participation in exercise sessions, patients who attended the orientation without subsequent exercise training still received 1.5 hours of educational information that addressed important material pertaining to exercise, nutrition, and lifestyle change. Third, loss of contact (30/47, or 64%) after hospital discharge was the most common reason for not attending orientation, despite confirmation of contact information by research staff. Consequently, a strategy of making appointments for CR after hospital discharge has shortcomings, and better strategies are needed to ensure continuing contact. Finally, among the 22 patients who attended the CR orientation session but decided not to participate in the exercise training sessions, 36% (8/22) cited insurance barriers as the most important reason for not participating in exercise, a problem that could be addressed by insurance systems and payers. Furthermore, any substantial delay created by the need to obtain insurance preauthorization is likely to decrease enrollment into CR.
When put in context, the results of the present study, in combination with the work of others,14,15 suggest that national guidelines11,12 should emphasize that an early appointment is preferable in medically stable patients and that long delays should be avoided. In addition, staff for each program should evaluate their program’s time interval between hospital discharge and CR enrollment. Any discovered obstacles that create unnecessary delay should then be removed. In short, an early appointment to CR, barring any individual patient safety concerns, should become the national standard of care.
Study limitations may include the absence of blinding of the recruiters and CR staff and the subsequent potential influence on patient outcomes; however, such blinding was not possible based on staff need to access the medical record for patient care. Second, this trial was conducted in a single urban hospital and may not be generalizable to other centers; however, we successfully coordinated among 5 CR centers in our metropolitan area, and our population was racially and socially diverse. Third, we were unable to approach all sequential eligible patients because of study resource limitations, yet our groups were generally balanced and reasonably representative of populations seen in CR. Fourth, we did not include surgical patients in the present study, and our results cannot be generalized to this patient group. Fifth, the unplanned interim analysis might weaken the statistical significance of the primary outcome; however, the visible difference in time to orientation demonstrates that the intervention did have an impact.
Although these limitations are notable, it is important to remember that the present study was prospective, randomized, and balanced between groups and met its primary end point. Regardless of group, we also demonstrated an overall attendance rate to the first contact with CR of 68% (101/148) and a rate of participation in ≥1 exercise session of CR of 53% (79/148), which is appreciably better than the current national average. In addition, there was a low likelihood of participation bias, which is a major limitation of most other studies in CR.
We demonstrated that early appointments to CR within 10 days of hospital discharge (versus 35 days) are feasible and significantly improve participation in the free-of-charge orientation session held before exercise training is begun. If such a system were implemented in other hospitals across the United States, it would likely have a favorable effect on initial participation in CR. Furthermore, it is unlikely to require special resources beyond those already in place in hospitals across the United States and would help to address an important performance gap10 in the care of the patients with established cardiovascular disease.
We acknowledge the help of Mohanad Alfaqih, MD, Mahmoud Assaad, MD, Moustafa El-Rafail, MD, Deirdre Mattina, MD, Ravish Parekh, MD, Michael Rothberg, MD, Sunay Shah, MD, and Diego Vasquez, MD, who each recruited patients for this study. We also express appreciation to all the staff in the Henry Ford Preventive Cardiology Unit who helped make this trial possible.
Sources of Funding
Funding for statistical analysis came from the Department of Graduate Medical Education at Henry Ford Hospital.
Dr Keteyian, Dr Ehrman, Brooks A. Hibner, and Melissa A. Vanzant each participate in a research grant from Janssen Healthcare Innovation, but this grant did not support this clinical trial. Dr Keteyian also serves as a consultant for Janssen Healthcare Innovation. The other authors report no conflicts.
- Received June 7, 2012.
- Accepted November 20, 2012.
- © 2013 American Heart Association, Inc.
- Giannuzzi P,
- Temporelli PL,
- Marchioli R,
- Maggioni AP,
- Balestroni G,
- Ceci V,
- Chieffo C,
- Gattone M,
- Griffo R,
- Schweiger C,
- Tavazzi L,
- Urbinati S,
- Valagussa F,
- Vanuzzo D
- Goel K,
- Lennon RJ,
- Tilbury RT,
- Squires RW,
- Thomas RJ
- 5.↵Receipt of outpatient cardiac rehabilitation among heart attack survivors: United States, 2005. MMWR Morb Mortal Wkly Rep. 2008;57:89–94.
- Suaya JA,
- Shepard DS,
- Normand SL,
- Ades PA,
- Prottas J,
- Stason WB
- Balady GJ,
- Ades PA,
- Bittner VA,
- Franklin BA,
- Gordon NF,
- Thomas RJ,
- Tomaselli GF,
- Yancy CW
- Arena R,
- Williams M,
- Forman DE,
- Cahalin LP,
- Coke L,
- Myers J,
- Hamm L,
- Kris-Etherton P,
- Humphrey R,
- Bittner V,
- Lavie CJ
- Thomas RJ,
- King M,
- Lui K,
- Oldridge N,
- Pina IL,
- Spertus J
- Williams M
- 18.↵American College of Sports Medicine. ACSM’s guidelines for exercise testing and prescription. 8th ed. Baltimore, MD: Lippincott Williams & Wilkins; 2010.
- Grace SL,
- Chessex C,
- Arthur H,
- Chan S,
- Cyr C,
- Dafoe W,
- Juneau M,
- Oh P,
- Suskin N
Outpatient cardiac rehabilitation (CR) is a highly effective but significantly underutilized intervention for patients after an acute myocardial infarction, percutaneous coronary intervention, or cardiac surgery. The mortality reduction from CR is similar to that seen with statin medications (range 20%–35%), yet only ≈30% of eligible patients participate in CR. National organizations, including the American Heart Association, have called for improvement of this important quality gap. One contributing modifiable factor may be the timing of the first appointment for CR. Recent retrospective studies have suggested that for each day of delay between hospital discharge and CR orientation, there is a 1% loss in CR participation. We prospectively randomized 148 patients to receive either an early or standard appointment to CR. Attendance occurred at a median of 8 versus 42 days per group, respectively. We found a significant 18% absolute and 56% relative improvement in attendance at orientation in the early group. Although the effects of an early appointment were limited to attendance at orientation, this technique was simple, safe, and was easily incorporated into the referral routine. Consequently, when coupled with other known effective techniques (automatic-referral and liaison-facilitated referral), an early appointment could potentially increase patient participation in CR and reduce the size of this important quality gap in the secondary prevention of heart disease.